CT Perfusion In Acute Ischemic Stroke

Aug 26, 2023

CT Perfusion In Acute Ischemic Stroke

00:00 - Intro
01:14 - Objectives
01:38 - Why CT perfusion?
04:23 - ASPECT scoring on non-contrast head CT
08:02 - Fundamental hemodynamic properties: CBF, CBV, MTT, Tmax
13:35 - Clinical uses: DEFUSE 3, DAWN, EXTEND
18:26 - Clinical examples
22:03 - Hypoperfusion index and multi-threshold Tmax maps
23:37 - Caveats and pitfalls: Caveats in estimating core
30:15 - Caveats and pitfalls: Caveats in estimating penumbra
35:37 - Summary
36:11 - Quality of study: Vessel selection, contrast opacification, patient motion
38:27 - Additional uses of CTP: Medium vessel occlusion
40:28 - Additional uses of CTP: Posterior circulation stroke
44:59 - Additional uses of CTP: Stroke mimics
46:52 - Can we use CTP like cardiologists use troponin?
47:58 - Summary and algorithm

Algorithm PDF may be found here: https://drive.google.com/file/d/1QLYj…

This video is intended for residents and fellows to understand the interpretation and clinical applications of CT perfusion imaging for the treatment of patients with acute ischemic stroke.

Created, produced, and narrated by:
Igor Rybinnik MD
Neurology Clerkship Director
Rutgers Robert Wood Johnson Medical School

Content experts:
Raymond Mirasol MD, Roger Cheng MD
Division of Stroke and Neurocritical Care
Department of Neurology
Rutgers Robert Wood Johnson Medical School

Sudipta Roychowdhury, MD FACR
Clinical Professor of Radiology, Neurosurgery, and Neurology
Rutgers Robert Wood Johnson Medical School
Director of Interventional Neuroradiology
Director of Magnetic Resonance Imaging
Robert Wood Johnson University Hospital

References:
1. Saver JL, et al. JAMA 2016, 316(12):1279
2. Nomani AZ, et al. Neurology 2021, 97:e2079-87
3. Sarraj A, et al. Stroke 2021, 52(1):57-69
4. Barber PA, et al. Lancet 2000;355(9216):1670-4
5. Almallouhi AZ, et al. JAMA Network Open. 2021;4(12):e2137708
6. Adapted from Sotoudeh H, et al. Acad Radiol 2019;26(11):1565-79
7. Albers GW, et al. NEJM 2018;378:708-718
8. Nogueira R, et al. NEJM 2018(1);378:11-21.
9. Ma H, et al. NEJM 2019;380(19):1795-1803
10. MacLellan A, et al. JSCVD 2022;31(1):106-208
11. Normani AZ, et al. Neurology 2021;97(21):e2079-87
12. Albers GW, Stroke 2018;49:1-5
13. Martins N, et al. Intervent Neurol 2018;7:513-521
14. Campbell BC, et al. Int J Stroke 2015;10:51–54
15. Amukotuwa SA, et al. Stroke 2021;52:3308-3317
16. Katyal A, et al. Eur J Neurol 2021;28:2657-2668
17. Tao WD, et al. Stroke 2012;43(8):2060-2065
18. Capasso R, et al. Neuroradiology 2021;63:317–330
19. Novakovic-White R, et al. Neurology 2021;97:S158-S169
20. Vagal A, et al. Neurology 2019;93(20):888-898.
21. Benson JC, et al. AJNR 2016;37:2239-44

Images adapted from:
- Articles referenced above
- Adobe Creative Cloud
- Blumenfeld H. Neuroanatomy Through Clinical Cases, 2nd ed. Sinauer, 2010.

Music:
- The Itch (Instrumental) - NEFFEX
- Icelandic Arpeggios - DivKid
- Under Cover - Wayne Jones

Disclaimer: Please note that this material was simplified for educational purposes. For patient management, please review your clinical society’s guidelines and engage expert consultation where appropriate. Also, the opinions of the presenters do not necessarily reflect those of Rutgers Robert Wood Johnson Medical School, Robert Wood Johnson University Hospital, RWJBarnabus Health, or Rutgers University as a whole.

Content

7.04 -> Welcome back, neurophiles.

8.4 -> Dr. Rybinnik here.

9.68 -> Today’s topic is interpreting CT perfusion

11.84 -> for the treatment of acute ischemic stroke.

14.32 -> This talk is a bit long and complicated,

16.16 -> but then perfusion is a complicated topic.

18.32 -> So, feel free to use chapter markers below,

20.56 -> slow things down,

21.52 -> speed things up.

22.4 -> Whatever helps you learn.

26.72 -> This video is rated R, for “Resident” due to clinical uncertainly, complex language

29.68 -> and moderate quality of evidence.

30.8 -> Medical students require accompanying resident or attending guardian.

35.92 -> I have no relevant financial conflicts to disclose.

39.76 -> As always,

40.4 -> the cases presented in this talk were modified slightly for teaching purposes.

45.6 -> CT perfusion is an exciting and rapidly evolving imaging modality,

49.12 -> but for the purpose of this talk,

50.56 -> we will try to adhere closely to high-quality evidence.

53.76 -> Debates about the finer points of perfusion

55.92 -> are outside the scope of this video.

58.88 -> Lastly,

59.6 -> no single imaging modality can qualify a patient for stroke intervention.

63.76 -> Let me say that again.

65.6 -> No single imaging modality can qualify a patient for stroke intervention.

70.24 -> You must always look at the entire clinical picture.

74.96 -> Why do we need CT perfusion?

76.8 -> Hopefully that will become clear in the first section.

79.76 -> We will then discuss the fundamental hemodynamic properties,

83.68 -> review well-validated clinical uses of CT perfusion,

87.52 -> discuss some caveats and limitations,

90.48 -> and conclude with some additional uses of this imaging tool,

93.04 -> and a tidy algorithm to bring it all together.

98.8 -> For decades,

99.68 -> we have used a non-contrast head CT

101.76 -> and CT angiogram to select patients for reperfusion therapy.

105.68 -> Why then do we need CT perfusion all of a sudden?

109.28 -> In 2015, multiple high-quality trials established thrombectomy as the new standard treatment

114.24 -> for acute ischemic stroke caused by large vessel occlusion.

117.6 -> These trials evaluated endovascular therapy within the standard window

121.36 -> (0-6 hours from stroke onset).

124.64 -> HERMES collaboration pooled data from these major trials

127.44 -> and analyzed how time from symptom onset to start of intervention

131.36 -> affects outcome.

134.32 -> The longer it takes to intervene on a patient with large vessel occlusion,

138.24 -> the lower the chance of favorable outcome.

141.84 -> And by about 7 hours from stroke onset,

144.48 -> the benefit of endovascular therapy

146.8 -> disappears.

149.36 -> There is no getting around it.

150.56 -> Time is still brain!

151.84 -> Time still matters when it comes to acute stroke interventions.

155.68 -> But there are some patients that can still benefit from intervention

159.04 -> beyond the conventional time windows.

162 -> Hypothetical scenario:

163.28 -> Say, you have 3 patients

164.56 -> with left middle cerebral artery occlusion.

166.96 -> At the time of occlusion,

167.84 -> the supplied brain is hypoperfused.

170.4 -> The entire territory is penumbra.

173.12 -> Without blood flow,

174.48 -> this hypoperfused brain

175.92 -> shifts to permanent infarction over time.

179.6 -> At 6 hours from occlusion,

181.12 -> in the first patient,

182 -> the entire MCA territory may have progressed to ischemic core.

186.64 -> In the second patient,

187.92 -> only half of MCA territory

189.76 -> progresses to irreversible ischemia.

192.96 -> And the third patient is still doing quite well,

195.12 -> with a very small ischemic core.

197.76 -> By 12 hours from the onset of occlusion,

199.92 -> patient 2 may look like patient 1

201.68 -> with completed MCA stroke.

204.24 -> But even at 24 hours,

206.4 -> the third patient may still have much viable brain.

209.44 -> This is the patient that may still benefit from reperfusion.

213.2 -> By the way,

213.68 -> the rate of stroke progression to ischemic core is highly variable

217.04 -> and depends on initial stroke size,

219.28 -> collaterals,

219.92 -> and age of the patient

221.2 -> (among other factors).

224.96 -> Let’s make this example more concrete.

227.84 -> A 46-year-old man with multiple vascular risk factors

230.64 -> presents with a left MCA occlusion.

233.28 -> Time from last known well is 1.5 hours.

236.64 -> His deficits are quite disabling

237.96 -> -- right hemiparesis, left gaze deviation, right field cut, significant aphasia

242.48 -> (what you would expect for a left MCA syndrome).

245.68 -> NIH stroke scale score is 14.

249.52 -> HERMES data tells us that within the standard thrombectomy window

252.8 -> (6 hours from stroke onset),

254.32 -> chances are there is significant amount of viable brain,

257.76 -> and we can use the good old non-contrast head CT

260.48 -> to select patients for intervention.

263.2 -> To assess ischemic core systematically,

265.12 -> let’s apply the ASPECTS,

266.8 -> which stands for

267.44 -> Alberta Stroke Program Early CT Score.

270.56 -> This scale divides the MCA territory

272.64 -> into 10 segments:

274.4 -> Head of caudate,

275.68 -> internal capsule,

277.04 -> lentiform nucleus,

278.64 -> insula,

279.92 -> and the 6 cortical segments.

282.88 -> A point is subtracted for each segment that is hypodense.

286.48 -> For example,

287.36 -> a patient with a completed deep MCA stroke

290.16 -> would lose 4 points,

292.08 -> and hypodensity in the cortical MCA territory could subtract

295.44 -> up to another 6 points.

300.32 -> What is this patient’s score?

304.16 -> Yep, this is a normal head CT.

305.92 -> So, the ASPECT score is 10.

309.2 -> Ok.

310.32 -> What if we change the timing?

312.72 -> Now we have a 67-year-old man with multiple vascular risk factors

316.08 -> who presents with a left MCA occlusion

318.08 -> just like in the last case.

320.48 -> But this time,

321.36 -> the patient woke up with deficits.

324.08 -> Time from last known well is 11 hours.

327.44 -> He has a full MCA syndrome,

328.88 -> and NIH stroke scale score is 20.

333.28 -> What is the ASPECTS in this case?

335.52 -> I’ll give you a minute.

340 -> If you find it difficult to call,

341.44 -> you are not the only one.

343.28 -> If you find it obvious,

344.56 -> don’t get too excited patting yourself on the back.

346.72 -> I adjusted the contrast for the purpose of this video,

348.96 -> so that the hypodensity projects better.

351.52 -> And there is a hypodensity

353.52 -> in the left head of caudate,

354.8 -> internal capsule,

355.6 -> and lentiform nucleus.

357.44 -> So, I would say that the ASPECTS is 7.

361.28 -> We certainly had to work hard to calculate it.

363.92 -> I think we can all agree that it is difficult to define ischemic core

367.04 -> on a non-contrast head CT.

370.24 -> Low ASPECTS supposedly correlates with a large core infarction.

373.6 -> But it is not even clear how well low ASPECTS predicts poor outcome.

377.84 -> For example,

379.04 -> in the large multicenter retrospective Stroke Thrombectomy and Aneurysm Registry

383.84 -> (STAR registry),

385.12 -> which was comprised of over 2000 patients

387.44 -> with internal carotid or M1 occlusion

389.52 -> presenting in the extended window,

391.36 -> patients with ASPECTS score of 2 to 5

393.52 -> (that’s up to 80% of the MCA territory supposedly damaged)

397.28 -> had 4 times higher chance of functional independence at 90 days

400.64 -> if they received mechanical thrombectomy.

403.76 -> What!?!

405.36 -> Come on!

406.16 -> I know it’s a retrospective cohort, but

408.72 -> good outcome with low ASPECTS!?

411.2 -> Clearly, we need a more reliable imaging modality

413.68 -> to select patients for interventions.

416.48 -> MRI is obviously the gold standard.

418.88 -> Diffusion-weighted imaging (DWI) abnormalities

422.08 -> reflect actual histological changes

424 -> that occur as a consequence of cerebral ischemia.

426.64 -> But MRI is slow and cumbersome.

428.72 -> Try getting a disoriented patient with a large vessel occlusion

431.6 -> to stay still for 30 minutes inside the MRI machine.

435.68 -> I am being a bit facetious.

437.76 -> Most major stroke centers do have protocols to get MRI quicker,

441.04 -> but CTP is more still prevalent

443.12 -> and only takes a few minutes.

447.28 -> This is a CT perfusion for the patient with just discussed.

450.56 -> An automated algorithm calculates and color-codes the ischemic core

454.4 -> and penumbra.

455.76 -> I bet I don’t even have to explain the numbers at the bottom of the page

458.64 -> for you to guess that the core is in pink,

460.56 -> and penumbra is in green.

464.4 -> And, there is a huge mismatch between the two.

468.72 -> Got it?

469.6 -> Awesome!

470.16 -> The talk is over.

470.96 -> See you later.

472.88 -> No, no, no.

473.68 -> That's not how we work around here.

474.88 -> Of course we will explain it.

476.48 -> Let’s dive in and briefly review the fundamentals first.

482.24 -> The mismatch map you just saw

483.84 -> is generated from the source images.

485.6 -> These are called image columns.

487.44 -> The first column are baseline CT images,

489.76 -> followed by four useful parameters:

493.28 -> Cerebral blood volume

494.8 -> is defined as the volume of blood flowing in 100 grams of brain tissue.

499.28 -> It’s an indirect measure of collateralization and autoregulation

502.48 -> -- the amount of blood that can be delivered to brain tissue.

507.92 -> Mean transit time

509.28 -> is the average time

510.24 -> it takes for blood to travel from intracranial arteries

513.04 -> to veins

513.92 -> across a capillary bed

515.76 -> (4 to 5 seconds is considered normal).

519.84 -> Cerebral blood flow

521.6 -> is the cerebral blood volume

523.52 -> (in 100 grams of brain tissue)

525.44 -> per minute.

527.04 -> Cerebral blood flow

528.32 -> is essentially cerebral blood volume

530.64 -> divided by the mean transit time

532.8 -> (the time it takes to deliver that volume).

535.6 -> Cerebral blood flow predicts the core infarction,

538.08 -> and does so better than cerebral blood volume.

541.84 -> Notice that CBV and CBF are preceded by a letter “r” for “relative.”

547.12 -> CT perfusion does not actually measure

549.2 -> the absolute volume or flow.

552.32 -> Instead,

553.04 -> blood flow and blood volume are expressed as a percentage

556.08 -> of a presumed healthy brain tissue.

558.16 -> So, in each patient,

559.2 -> an ischemic lesion in one hemisphere

561.28 -> may be compared to healthy tissue in the opposite hemisphere.

567.12 -> The last parameter,

568.16 -> time-to-Maximum, or Tmax,

570.24 -> represents the time from the start of the scan

572.64 -> until the maximum intensity of contrast bolus arrives into the brain.

578.88 -> Delay in Tmax correlates with penumbra.

583.76 -> For those of you who enjoy graphs,

585.76 -> if you plot the density

587.04 -> (or brightness)

587.76 -> of injected contrast bolus versus time,

590.32 -> you will get a graph like this.

595.28 -> Tmax is the timepoint at which contrast peaks intracranially.

600.88 -> Mean transit time

601.84 -> is the average time

603.2 -> that contrast spends inside the intracranial blood vessels.

609.28 -> Cerebral blood volume

610.4 -> is the area under the curve.

612.88 -> and cerebral blood flow is related to

614.72 -> how quickly that blood volume gets to the brain tissue

617.84 -> -- the slope of the upstroke.

622.32 -> In a patient with early large vessel occlusion,

624.56 -> arrival of the blood to brain tissue is delayed

630.64 -> (blood flow takes the side streets or collaterals

632.8 -> instead of the main highway).

637.28 -> But neurons are still ultimately getting the same blood volume,

641.2 -> at least until ischemia sets in.

645.92 -> Since cerebral blood flow

646.96 -> is cerebral blood volume divided by mean transit time,

650.32 -> when mean transit time increases,

653.12 -> cerebral blood flow decreases.

657.28 -> Again,

657.6 -> CT perfusion does not actually measure absolute blood flow

660.88 -> or ischemic changes like MRI does.

663.2 -> We use CTP abnormalities as a surrogate marker

666.16 -> of tissue hypoperfusion.

668.24 -> This will become important when we talk about CTP’s pitfalls

671.04 -> and limitations.

674.8 -> Back to the image columns.

676.48 -> Note the color scale.

677.84 -> Lower values are in blue,

679.68 -> intermediate values are in green-yellow,

681.84 -> and high values are in red.

686.08 -> Let’s take a look at our patient with left MCA occlusion.

689.6 -> Which parameter is most obviously abnormal?

693.28 -> Yes!

694.08 -> Tmax is clearly delayed

696 -> (colors are more yellow-red).

698.32 -> Tmax is great at identifying penumbra

700.72 -> because it reveals high contrast between areas of delay

704.16 -> and normal tissue.

706 -> Tmax does not vary with tissue type like cerebral blood volume does

710 -> (grey and white matter receive different blood volumes),

712.88 -> and this uniformity produces flat contrast in normal brain.

716.72 -> So even small or subtle areas of Tmax delay

719.84 -> are highly visible.

722.8 -> Mean transit time is also increased

725.04 -> as expected.

727.04 -> The decrease in cerebral blood volume in the subcortical left MCA territory

731.28 -> (the bluer colors on the patient’s left),

733.36 -> represents infarcted tissue.

736.32 -> And since cerebral blood flow

738 -> is cerebral blood volume divided by mean transit time,

741.6 -> cerebral blood flow is also decreased.

745.76 -> Interpreting these perfusion maps

747.28 -> is like reading your fortune in tea leaves.

749.2 -> I get it, I get it.

750.56 -> So for more consistent and objective results,

752.88 -> these perfusion maps are usually subjected to some form of

756.24 -> thresholding.

757.36 -> We don’t want to see everything all at once.

759.36 -> Just the important bits.

761.28 -> And that’s how we end up with this final image.

764.48 -> On the left,

765.6 -> is the volume of a region with cerebral blood flow less than 30%

768.8 -> of patient’s normal.

770.48 -> CBF less than 30% cutoff is widely validated in clinical trials.

775.44 -> On the right,

776.64 -> is an area of Tmax with a delay more than 6 seconds.

780.32 -> This metric provides a reasonable estimate of final infarction

784 -> in patients without reperfusion.

785.92 -> So Tmax is used as a surrogate for penumbra.

789.68 -> Subtract this patient’s volume of the ischemic core

792.48 -> (13 mL)

793.68 -> from the volume of penumbra

795.44 -> (110 mL),

796.8 -> and we arrive at the final mismatch volume of 97 mL.

800.8 -> The difference between ischemic core and penumbra

802.8 -> can also be represented as a ratio.

805.36 -> In other words,

806.08 -> our patient has 8.5 times more penumbra

809.6 -> than core.

811.6 -> Let’s see how this works in clinical practice.

816.56 -> Our clinically relevant cutoffs for stroke intervention come from three landmark trials:

821.52 -> DEFUSE 3, EXTEND, and DAWN.

827.2 -> DEFUSE 3 showed that patients with an ischemic core of at most 70cc

832.64 -> (about one third to one half of MCA territory),

835.84 -> and

836.48 -> a potentially salvageable penumbra of at least 1.8 times that of the core

841.6 -> have better outcomes with thrombectomy.

846.16 -> The difference between the core and penumbra

847.84 -> was also expressed as an absolute volume of potentially reversible ischemia

851.84 -> of at least 15ccs.

857.76 -> EXTEND trial

858.8 -> extended the window for intravenous TPA

861.2 -> to 9 hours.

862.96 -> That trial selected patients with similar core size as DEFUSE 3,

867.44 -> but allowed a smaller core-to-penumbra mismatch

870 -> of at least 1.2 times

872.96 -> (or least 10cc absolute difference).

876.88 -> By the way,

877.6 -> large vessel occlusion was not required in the EXTEND trial.

880.96 -> But by definitions,

882.64 -> you can see that some patients that meet DEFUSE 3 criteria,

886.56 -> may also meet the EXTEND criteria.

892.32 -> The DAWN trial had the most complicated inclusion criteria.

896.32 -> Since size of the completed stroke

898.24 -> is arguably one of the most important predictors of clinical outcome,

901.84 -> that trial focused on selecting a group of patients

904.48 -> with smaller ischemic cores

907.6 -> and a large clinical deficits.

912.4 -> Did you catch that?

914.48 -> There was no penumbral requirement.

917.12 -> As long as patients had small core infarctions,

919.92 -> clinical deficits were used as a surrogate

922.08 -> for potentially reversible ischemia.

926 -> For the very elderly (over the age of 80),

929.36 -> intervention was offered to patients with a small core

932.48 -> (at most 20 cc),

934.32 -> and moderate-to-severe clinical deficits

936.56 -> (NIH stroke scale score of at least 10).

939.76 -> For patients under 80 years of age,

942.4 -> a core of at most 30 cc was allowed with moderate deficits

946.24 -> (NIH stroke scale score of 10-19),

951.12 -> and a core of up to 50 cc

953.28 -> was allowed if the deficits were severe

956.24 -> (NIHSS score of at least 20).

962.96 -> In the DEFUSE 3 trial,

964.16 -> patients who received thrombectomy

965.76 -> presented with a proximal large vessel occlusion

968.48 -> within 6 to 16 hours from last known well.

971.92 -> Median age was 70,

973.28 -> give or take 11 years.

975.52 -> Deficits were severe

976.88 -> with median NIHSS score of 16.

980 -> Most (about two thirds) had unwitnessed stroke onset,

983.2 -> usually awakening with stroke symptoms.

985.68 -> While MRI was allowed was patient selection,

987.92 -> majority of the patients were selected by CT perfusion.

991.76 -> Median ischemic core was 9cc,

994.24 -> far less than the maximum of 70cc's.

997.44 -> And median penumbra was 115cc,

1000.24 -> with a mismatch of over 10 times,

1002.8 -> again, far exceeding the prespecified ratio of 1.8.

1007.84 -> DEFUSE 3 was terminated early

1010.4 -> for efficacy

1011.68 -> after only 182 patients were enrolled.

1016.16 -> DAWN trial offered thrombectomy to patients with a proximal large vessel occlusion

1021.44 -> presenting within 6 to 24 hours from last known well.

1025.6 -> Median age was and deficit severity

1027.2 -> were similar to the DEFUSE 3 patient population.

1030.32 -> Vast majority of patients in the DAWN trial had unwitnessed stroke onset.

1035.44 -> Just like in DEFUSE 3, most patients were selected by CTP.

1040.8 -> And the median ischemic core was small (8cc).

1046.48 -> EXTEND focused on intravenous TPA,

1048.88 -> enlarging the window from four and half to 9 hours

1052.4 -> after last known well.

1054 -> By the way, if patients woke up with stroke symptoms,

1057.2 -> the onset of stroke was defined to be the midpoint of sleep.

1061.04 -> Technically, if a patient went to sleep at 11pm

1063.68 -> and woke up at 9am with stroke symptoms,

1066.24 -> as far as the EXTEND trial is concerned,

1068 -> stroke onset would be 5 hours ago.

1071.36 -> EXTEND patient population was aged similarly to DEFUSE 3 and DAWN populations,

1075.68 -> but had milder strokes

1077.6 -> (median NIH stroke scale score of 12).

1080.64 -> Majority of strokes were unwitnessed.

1083.12 -> Most patients were selected by CTP.

1086 -> And ischemic cores were small.

1088 -> Median penumbra was 74cc,

1090.32 -> with a mismatch of nearly 15 times,

1093.36 -> far exceeding the prespecified ratio of 1.2.

1097.36 -> Also, large vessel occlusion was not a requirement for EXTEND,

1100.56 -> but nearly 70% of patients ended up having occlusions.

1106.88 -> Take one last look at the criteria, because we are about to apply them to real cases.

1114.24 -> Let’s get back to the patient with whom I started the discussion of CT perfusion.

1118.16 -> As a reminder,

1118.88 -> he is 67-year-old man

1120.24 -> with multiple vascular risk factors

1122.48 -> who presented within 11 hours from last known well

1125.36 -> with global aphasia, right hemianopia, left gaze deviation, and right hemiparesis.

1130.32 -> NIHSS score was 20.

1133.04 -> Left M1 occlusion was confirmed with CTA.

1137.2 -> Would you offer thrombectomy?

1143.52 -> Let’s apply DEFUSE 3 criteria.

1146.16 -> Ischemic core is only 13cc,

1148.4 -> which is less than the pre-specified 70cc.

1151.76 -> And the core-to-penumbra mismatch is 8.5 times,

1155.2 -> which is certainly higher than the prespecified 1.8 times.

1159.12 -> So, this patient will benefit from thrombectomy.

1166.08 -> Here is another case:

1167.44 -> A 74-year-old man with atrial fibrillation but off anticoagulation.

1172 -> He presented 7 hours from midpoint of sleep

1174.96 -> (11 hours from last known well).

1177.2 -> Right A2 was occluded.

1179.6 -> He has left sensorimotor deficits, paucity of speech, and dysarthria.

1183.52 -> NIH stroke scale score is 11.

1187.04 -> Would you offer reperfusion therapy?

1194.16 -> We can apply EXTEND criteria here.

1196.72 -> Mismatch is more than 1.2 times,

1199.44 -> and the ischemic core is

1203.28 -> so intravenous TPA may be considered.

1206.24 -> Technically DEFUSE 3 criteria would apply here as well,

1209.68 -> but the lesion is too distal for thrombectomy

1212.8 -> (at least according to our current evidence).

1217.36 -> Ok.

1218.24 -> What about this patient?

1219.84 -> 94-year-old woman with atrial fibrillation,

1222.16 -> who presented 12 hours from last known well

1224.8 -> with acute left M1 occlusion

1226.96 -> causing encephalopathy, left gaze deviation, right hemiparesis, right field cut, and global aphasia.

1232.56 -> NIH stroke scale score is 18.

1235.92 -> To intervene or not to intervene?

1242.88 -> Let’s apply DEFUSE 3 criteria.

1244.88 -> Mismatch is favorable

1246.24 -> (more than 1.8 times),

1248.48 -> but the ischemic core is over 70cc.

1253.44 -> Risk/benefit profile for a thrombectomy in this patient is likely unfavorable.

1261.28 -> And the last example.

1263.2 -> 48-year-old man with dyslipidemia

1265.28 -> presents six and a half hours from last known well.

1268.56 -> He went to sleep and awakened with severe deficits

1271.12 -> including disorientation, right gaze

1272.88 -> preference, left sensorimotor deficits, dysarthria, and neglect.

1276.8 -> NIH Stroke Scale score is 19.

1279.92 -> Right M1 occlusion is confirmed.

1282.96 -> Here is his perfusion imaging.

1284.96 -> Would you recommend thrombectomy for this patient?

1289.52 -> Of course!

1290.56 -> DEFUSE 3 criteria are satisfied here.

1293.28 -> Patient is within 6-16 hours from last known well.

1296.56 -> There is a large vessel occlusion.

1298.4 -> Ischemic core is 24cc,

1300.4 -> which is less than 70,

1302.08 -> and penumbra is 4 times larger than the core.

1308.32 -> And speaking of penumbra,

1310.24 -> can we predict if this penumbra is about to convert to ischemic core?

1314.32 -> What are the chances of this infarct growing?

1319.12 -> For that,

1319.84 -> we need to discuss different Tmax thresholds.

1324.96 -> Tmax of more than 6 seconds (in green) is how we typically define penumbra.

1329.6 -> Tmax of more than 10 seconds (in red)

1333.04 -> estimates the tissue at immediate risk of infarction.

1337.04 -> Dividing the volumes defined by these two delay thresholds,

1339.84 -> gives you the hypoperfusion index

1342.08 -> (also known as the hypoperfusion intensity ratio or HIR).

1346.88 -> Hypoperfusion index of over 0.34 in the DEFUSE 3 cohort

1352.56 -> predicted poor collateral flow and significant infarct growth

1357.12 -> early over the first 24 hours.

1360.24 -> In this patient, we expect the ischemic core to grow rapidly without treatment.

1367.52 -> By contrast, the 4-6 second delays are mild enough

1370.56 -> that the tissue is unlikely to infarct.

1373.76 -> By the way,

1374.16 -> in a nice bit of synergy,

1375.36 -> the colors on this multi-threshold map correspond to the colors on the original image columns

1380.08 -> without any thresholds applied.

1385.36 -> So riddle me this:

1387.52 -> Based on these Tmax multi-threshold maps,

1389.92 -> which of these two patients with left MCA occlusion

1392.4 -> is more likely to experience ischemic core growth?

1400.64 -> Yes, this one.

1403.52 -> Sadly, he completed his MCA infarction

1406.16 -> and needed decompressive hemicraniectomy to treat malignant cerebral edema.

1413.12 -> I hope by now you have realized

1414.56 -> that CT perfusion is incredibly useful.

1416.96 -> But just like any tool,

1418.4 -> it is not without limits.

1421.12 -> We determine candidacy for intervention based on estimation of ischemic core and penumbra.

1427.36 -> But, what if those estimations are inaccurate?

1430.48 -> Are there situations where CTP can lead us astray?

1434.08 -> By the tone of my question,

1435.2 -> the answer of course is “yes.”

1437.04 -> Let’s take ischemic core for example.

1439.12 -> Cerebral blood flow measurement may under or overestimate the ischemic core.

1444.16 -> Let’s talk about underestimating the core first.

1448.48 -> Remember that CTP maps do not actually identify infarcted tissue.

1452.48 -> They identify regions with low blood flow that can predict infarcted tissue.

1457.12 -> But cerebral blood flow in a patient with a large vessel occlusion

1459.84 -> can change with time.

1462.4 -> Imagine a hypothetical patient

1463.84 -> with acute left middle cerebral artery occlusion.

1466.08 -> Cerebral blood flow is illustrated here as a function of time.

1469.84 -> Here is the well-accepted CBF less that 30% threshold.

1474.24 -> At 2 hours after stroke onset,

1476.32 -> cerebral blood flow drops significantly

1478.8 -> (below 30% of normal).

1481.36 -> Non-contrast head CT is still normal,

1483.92 -> and CBF maps register a core infarction

1486.72 -> (in pink).

1487.84 -> As expected, Tmax map show a sizable penumbra.

1492.72 -> By 8 hours,

1493.84 -> a hypodensity may be now visible on CT,

1496.48 -> corresponding to a larger core infarction on the CBF map.

1500.88 -> But at the extreme of the 24-hour time window,

1503.28 -> leptomeningeal collaterals may have already been recruited,

1506.64 -> and cerebral blood flow may improve a bit

1509.52 -> just above the 30% threshold.

1512.96 -> So, core infarction “disappears” from the CBF map,

1516.16 -> while Tmax still shows a significantly delayed flow in the left MCA territory.

1520.88 -> This is a situation where collaterals came in

1523.2 -> “too little, too late.”

1524.56 -> It doesn’t matter that blood flow is better.

1526.72 -> The tissue is already irreversibly injured.

1530.64 -> In this case, only the non-contrast head CT

1533.84 -> (or diffusion weighted imaging MRI)

1535.92 -> can show you the true ischemic core.

1538.8 -> Practically, this discussion is relevant for two reasons.

1541.76 -> First,

1542.4 -> in a patient with unknown time of stroke onset,

1544.96 -> CT perfusion must be interpreted together

1547.68 -> with the non-contrast head CT and clinical history.

1552 -> Some software packages make this easier

1554.08 -> by overlaying the CT hypodensity on the CBF map.

1557.76 -> Here, an area in blue is a hypodensity threshold of 5-12 Hounsfield units,

1562.24 -> which is typical of acute infarcts.

1564.96 -> Some people even advocate relying on the gold standard MRI

1568.4 -> instead of the non-contrast head CT or CT perfusion

1571.76 -> the closer you get to the extreme of that 24-hour window.

1578.4 -> The other practical point

1580.08 -> is that CT perfusion maps are very time dependent!

1583.92 -> They only reflect the hemodynamics

1585.92 -> at the moment the scan is done.

1588.32 -> So, if the CTP is performed at a primary stroke center,

1591.28 -> and the patient is transferred to a thrombectomy-capable center

1594.72 -> by the time that patient arrives,

1596.88 -> the initial CTP may no longer be valid,

1599.52 -> especially if there is a significant transfer delay.

1602.48 -> So, you might have to repeat the imaging.

1607.76 -> A similar situation of underestimating core also occurs after thrombectomy.

1612.72 -> Cerebral blood flow improves,

1614.48 -> and the ischemic core disappears from CTP

1617.36 -> because there is now blood flow in the irreversibly

1619.84 -> injured region.

1621.2 -> But in this case,

1622.08 -> penumbra also disappears

1623.6 -> so you know you’ve made progress at improving blood flow delays.

1629.44 -> So, core may be underestimated when ischemic tissue is reperfused by thrombectomy

1633.6 -> or after leptomeningeal collaterals are recruited.

1636.88 -> Overestimation of the core is arguably a bigger problem

1640.4 -> because large core may exclude a patient from intervention.

1644.32 -> When a large vessel occludes,

1645.92 -> cerebral blood flow in that vessel’s territory dips below

1648.72 -> 30% of normal in fairly quickly.

1651.28 -> But, hypoperfusion does not necessarily translate to tissue ischemia.

1656 -> Not if it doesn’t last for long.

1658.8 -> If a patient presents early enough after stroke onset,

1661.68 -> and reperfusion is quickly accomplished,

1664 -> the core seen on CTP would disappear.

1666.8 -> This is especially true of patients

1668.88 -> presenting within the golden hour.

1671.68 -> By you will say,

1672.64 -> Dr. Rybinnik,

1673.76 -> in DAWN and DEFUSE 3 trials CTP was performed at or beyond 6 hours from last known well,

1678.96 -> so none of the patients were scanned within the golden hour.

1681.6 -> Weren’t they?

1685.04 -> Patients with unknown stroke onset

1686.8 -> were allowed in those trials.

1688.8 -> Let’s say a patient presents with symptoms on awakening

1691.52 -> (last known well was at bedtime,

1693.28 -> 9 hours ago).

1694.64 -> That patient comes to the emergency department

1696.8 -> and gets the following scan.

1699.12 -> Large core,

1700.32 -> large penumbra,

1701.52 -> no mismatch.

1702.96 -> Patient is not a candidate for intervention.

1705.36 -> Right?

1707.52 -> But what if

1708.72 -> it’s a situation like this

1710.16 -> – severe hypoperfusion but no ischemia.

1714.4 -> The non-contrast head CT provides a clue.

1717.36 -> Head CT is normal,

1718.48 -> suggesting that the patient may be within the early time window.

1722.72 -> This patient underwent thrombectomy

1724.16 -> for occlusion of the right internal carotid artery terminus,

1727.44 -> and contrary to the CTP,

1729.04 -> a follow-up CT on day 3 showed no large, completed infarction.

1734.4 -> The ischemic core in this case was overestimated

1737.04 -> – so-called “ghost” core.

1739.92 -> It didn’t really exist.

1742.8 -> The shorter the time between symptom onset to imaging

1746 -> and from imaging to reperfusion,

1748.16 -> the higher the probability of ischemic core overestimation.

1751.76 -> Thankfully, situations like this are rare

1753.84 -> (one in six patients in a retrospective cohort).

1757.04 -> If you think you are dealing with a patient within this very early time window,

1760.48 -> you may consider a more stringent CBF threshold

1763.36 -> of less than 20%.

1765.04 -> The difference in volume between the less than 20% and the accepted less than 30%

1769.6 -> cerebral blood flow threshold

1771.36 -> could be quite substantial.

1773.68 -> Look at this example.

1774.88 -> The ischemic core by accepted criteria is 44cc's.

1779.12 -> Now if this patient is over 80 years of age,

1781.52 -> they might be excluded from intervention

1783.2 -> based on the DAWN criteria.

1784.96 -> But applying the more stringent threshold for ischemic core

1788.48 -> removes the core completely

1790.32 -> and may allow us to intervene.

1795.6 -> So, let’s review:

1797.68 -> CBF may overestimate the core

1799.92 -> if CT perfusion is performed too early with respect to stroke onset.

1804.4 -> CBF may underestimate the core

1806.96 -> if CT perfusion is performed late with respect to stroke onset,

1810.24 -> or after reperfusion.

1814.96 -> Ok!

1816 -> Now let’s discuss overestimating penumbra.

1820.8 -> It is 4am,

1821.84 -> and you respond to a stroke code in cardiothoracic ICU.

1826.24 -> A code was called for a 63-year-old man with multiple vascular risk factors,

1830.4 -> including atrial fibrillation.

1832.32 -> He was off anticoagulation for a coronary artery bypass grafting surgery,

1836.56 -> which he underwent earlier today without any complications.

1840.96 -> When sedation was weaned,

1842.64 -> the nurse noted significant expressive aphasia, mild right

1845.52 -> hemiparesis, and dysarthria.

1848.24 -> NIH Stroke Scale score is 7.

1851.6 -> He was last known well 11 hours ago,

1854 -> pre-op.

1855.76 -> Non-contrast head CT,

1857.12 -> CTA and CT perfusion is performed.

1859.84 -> Head CT was normal.

1861.84 -> What’s your interpretation of this perfusion study?

1869.52 -> There are no areas of ischemic core

1871.28 -> (CBF less than 30%).

1873.84 -> Tmax is delayed beyond 6 seconds

1875.68 -> in the left ACA and MCA territories.

1879.44 -> Total mismatch is 115 cc.

1884 -> Seems to me that’s a perfect patient for intervention?

1888.08 -> But the CTA of the head shows preserved intracranial flow.

1891.12 -> There is some asymmetry between the two sides,

1893.12 -> but nothing is occluded.

1894.96 -> What’s going on?

1900.16 -> To the source images!

1902.8 -> Tmax is definitely increased in the left hemisphere.

1906.48 -> Mean transit time is barely increased.

1909.28 -> And cerebral blood flow and cerebral blood volume

1911.84 -> are essentially symmetric between the two hemispheres.

1917.36 -> Have you figured it out?

1921.36 -> Tmax,

1922.08 -> the maximum amount of time required for contrast to reach the brain tissue

1926.08 -> is sensitive to delay in flow.

1928.24 -> For example,

1928.88 -> if the internal carotid artery is chronically severely stenosed or occluded,

1933.12 -> Tmax will be increased.

1936 -> But a chronically occluded carotid is generally a well-compensated state.

1940.64 -> Mean transit time may be mildly increased or normal.

1945.44 -> More importantly,

1947.52 -> cerebral blood volume will be normal.

1950.88 -> That hemisphere is getting enough blood,

1953.2 -> even though it’s delayed,

1954.64 -> probably because it is arriving through collaterals.

1958.24 -> Don’t believe me?

1959.44 -> Well, thankfully

1960.32 -> perfusion software packages also provide multi-threshold maps

1963.52 -> for cerebral blood volume.

1965.44 -> In our patient,

1966.4 -> no matter what threshold is applied,

1968.24 -> cerebral blood volume seems to be preserved.

1972.72 -> CBV measures the relative volume of tissue that is vascularized

1976.48 -> (as compared to the unaffected side).

1978.72 -> This parameter generally does not care

1981.04 -> how long it would take for that tissue to receive that blood flow,

1984.08 -> as long as that blood flow eventually arrives.

1987.28 -> High CBV is a predictor of good collaterals,

1990.08 -> and low CBV can be a predictor of infarct core.

1993.92 -> (By the way,

1994.32 -> as I said previously,

1995.44 -> CBF seems to perform better at estimating core,

1998.48 -> and that’s why rely on CBF instead of CBV).

2002.96 -> So, cerebral blood volume is not time-dependent.

2007.52 -> Tmax is time dependent.

2010.48 -> If source maps show delayed Tmax with preserved CBV,

2014.48 -> you are probably dealing with a chronic carotid high-grade stenosis or occlusion,

2018.24 -> that's well-compensated by collaterals.

2020.72 -> It is in this situation,

2022.8 -> Tmax may overestimate the penumbra.

2026.88 -> That’s what this patient had.

2030.48 -> The lesion was emergently stented

2032.24 -> because of patient’s severe disabling deficits.

2035.12 -> In the interest of full disclosure,

2036.64 -> I should mention that this is individualized medicine.

2039.84 -> There is no high-quality evidence for something like this,

2042.32 -> and certainly not everyone with such pathology should be acutely stented.

2046.72 -> Case in point,

2047.52 -> here is another example.

2049.28 -> 62-year-old man

2050.4 -> with multiple vascular risk factors

2052.16 -> who noted slurred speech during breakfast

2054.56 -> (the first time he spoke after awakening).

2057.36 -> Last known well was 13 hours ago.

2060.08 -> He exhibits dysarthria, mild left facial droop and left arm drift.

2064.08 -> NIH Stroke Scale score is 3.

2067.36 -> CTA during the stroke code showed

2069.28 -> complete occlusion of the right cervical ICA,

2071.84 -> but intracranial flow was preserved.

2076.24 -> Interpret his CTP for me.

2081.28 -> Right!

2082.48 -> There is a small ischemic core in the right hemisphere.

2085.44 -> Once again, penumbra is large

2087.28 -> (152 mLs),

2089.04 -> and there is a significant mismatch of nearly 14 times.

2093.44 -> But now you know better than to fall for this.

2095.44 -> You are probably saying, “Show me the source images!”

2099.04 -> Ok!

2100.16 -> Tmax is significantly delayed

2102.24 -> (no surprise there).

2104.16 -> There is even some delay in mean transit time.

2108.08 -> But cerebral blood volume

2109.68 -> once again looks symmetric between the two hemispheres.

2112.88 -> CBV does not appear to be decreased enough to meet any of the accepted thresholds.

2118.24 -> Because of mild deficits,

2119.6 -> no intervention was pursued.

2121.2 -> This patient was treated medically with blood pressure augmentation.

2124.72 -> Patient’s deficits improved,

2126.64 -> and on repeat CT perfusion,

2128.24 -> the area of penumbra has decreased from 152mLs

2132.32 -> to only 32 mLs.

2137.68 -> Let’s review.

2139.36 -> Again.

2140.96 -> CBF may overestimate the core

2143.36 -> if CT perfusion is performed too early with respect to stroke onset.

2147.76 -> CBF may underestimate the core

2150.32 -> if CT perfusion is performed late with respect to stroke onset,

2153.84 -> or after reperfusion.

2156.4 -> Tmax is time-delay dependent

2158.56 -> and may overestimate the penumbra

2160.72 -> in patients with chronic high-grade stenosis or occlusion

2163.36 -> of the cervical carotid.

2170.56 -> The discussion we’ve been having until this point

2172.8 -> assumed a good quality study.

2175.28 -> Sometimes, you get images that do not make any sense.

2178.96 -> Then how do we determine that the study is technically adequate?

2183.12 -> In order to generate the perfusion maps,

2185.2 -> the software needs to identify blood going into and out of the brain

2188.96 -> by way of normal vessels.

2191.36 -> In this study,

2192.24 -> the algorithm selected right MCA as it’s arterial input function,

2196.56 -> and the straight sinus as it’s venous output function.

2200.08 -> Change in brightness is plotted versus time

2202.4 -> as the contrast bolus flows through the artery,

2204.8 -> and then exits the brain through the venous sinus.

2208.32 -> In a technically adequate study,

2210.4 -> the arterial input function needs to be bright enough,

2213.52 -> and usually exceeds a hyperdensity of 150 Hounsfield units

2219.68 -> The software also keeps track of side-to-side and rotational movement

2223.04 -> and generates plots for both in three dimensions.

2227.36 -> So, very little movement is present.

2230 -> Artery and vein are properly selected.

2232.4 -> Arterial input function is appropriately bright.

2235.2 -> This

2235.84 -> is a technically adequate study.

2239.28 -> Now, this image is technically inadequate.

2241.92 -> And that's an understatement.

2244.08 -> Patient moved and rolled during the scan, rendering arterial input function inadequate.

2249.84 -> So that’s movement.

2250.88 -> But issues with proper selection of vessels show up from time to time as well.

2255.04 -> Here, the algorithm focused on bone

2257.12 -> instead of an artery.

2258.64 -> Thankfully, the image can be reprocessed with correct vessel selection.

2263.28 -> And finally,

2264.24 -> the most common locations where penumbral measurement is overcalled

2267.92 -> are at the base of the skull

2269.28 -> and the orbits.

2270.96 -> Here,

2271.6 -> even minor motion and partial volumed bone

2274.64 -> create this artifactual “delayed” Tmax.

2280.4 -> Remember that the algorithm includes these areas

2283.04 -> when calculating the total volume of impaired perfusion.

2286.32 -> So,

2286.64 -> while these artifacts are easy to spot,

2288.96 -> they are actually pretty difficult to subtract from the final volumes.

2296.64 -> Ok.

2297.36 -> Those were the basics.

2299.2 -> Modern automated CT perfusion algorithms were designed for practicality.

2303.28 -> But, you can always extract more information

2305.52 -> when you know where to look.

2309.2 -> The best evidence for the use of CT perfusion in the treatment of acute ischemic stroke

2313.2 -> exists for large vessel occlusions.

2316.4 -> What about medium vessel occlusions?

2320.16 -> 62-year-old man with heart failure felt a shock of a defibrillator,

2323.68 -> and collapsed without losing consciousness.

2326.88 -> He arrived to the emergency department within the golden hour

2329.92 -> (50 minutes from symptom onset).

2333.36 -> He was noted to have left sensorimotor deficits and mild left neglect

2337.44 -> (NIH Stroke Scale score of 5).

2341.6 -> Here is his CTA of the head,

2343.52 -> axial maximal intensity projection.

2346.4 -> Is there a large vessel occlusion?

2353.36 -> Here, is the coronal maximal intensity projection.

2359.28 -> Challenging, isn’t it?

2362.96 -> CT perfusion was also done.

2365.04 -> The territory of Tmax over 6 seconds

2367.52 -> matches inferior right M2 division,

2370.16 -> and there is substantial mismatch.

2373.12 -> Reconstructing the CTA in the sagittal view finally allowed us to visualize

2377.12 -> the right M2 occlusion.

2379.12 -> This patient was treated with TPA at an outside hospital,

2382.4 -> and by the time he arrived at a comprehensive stroke center,

2385.12 -> his deficits improved nearly to baseline

2387.28 -> and thrombectomy was not offered.

2389.36 -> But the point here is that CTP helped identify

2392.56 -> the medium vessel occlusion.

2397.12 -> Even in experienced stroke centers,

2398.96 -> one in three distal vessel occlusions are missed with CTA alone.

2403.04 -> And interpretation of CTA is time-consuming.

2405.76 -> I mean, how many CTA reconstructions did we

2408 -> just have to review

2408.88 -> to catch that one medium vessel occlusion?

2411.76 -> Retrospective series suggest that

2413.52 -> adding Tmax helped catch virtually all of them.

2417.68 -> And as a bonus,

2418.8 -> when CT perfusion was available in these cases,

2421.44 -> CTA was cleared faster

2423.12 -> and with more confidence.

2427.68 -> So, CTP may help identify medium vessel occlusions.

2431.12 -> But what about posterior circulation strokes?

2434.8 -> Crash course on posterior circulation stroke:

2437.28 -> Majority of patients present with non-specific symptoms

2439.84 -> like dizziness, dysarthria and encephalopathy.

2443.2 -> Less than 5%

2444.8 -> have the signs with highest predictive value

2446.88 -> like Horner, crossed sensory deficits,

2448.8 -> quadrantanopia, oculomotor palsy, and motor deficits.

2452.8 -> So, imaging becomes key in the diagnosis.

2456.4 -> Non-contrast head CT

2459.2 -> misses over two thirds of acute posterior circulation strokes.

2463.12 -> Did you hear that,

2464 -> my future emergency medicine colleagues?

2466.4 -> I have very little confidence in the non-contrast head CT

2469.44 -> in these situations.

2472.08 -> Adding CTA roughly doubles the sensitivity.

2475.52 -> Incidentally, the sensitivity is similar to relying on CTP alone,

2479.44 -> but that is not the standard practice.

2481.6 -> Also,

2481.92 -> you have to be careful of those base of the skull artifacts.

2487.12 -> A combination of CT, CTA and CTP

2490 -> can help diagnose 9 out of 10 patients

2492.48 -> with posterior circulation stroke.

2494.8 -> Well, now we are talking!

2497.92 -> Consider this case.

2499.44 -> 86-year-old woman with atrial fibrillation on Apixaban,

2502.56 -> as well as other vascular risk factors

2504.72 -> who developed new vertigo, nausea, vomiting and dysarthria

2507.84 -> -- largely non-specific symptoms.

2511.04 -> Reportedly,

2511.68 -> she felt unwell for several days,

2513.44 -> but symptoms worsened within the past two hours.

2517.2 -> On arrival to the emergency department,

2519.12 -> her NIH Stroke Scale score was 2

2521.36 -> for mild ataxia.

2522.96 -> That’s not particularly surprising,

2524.56 -> since NIH Stroke Scale is really designed to quantify

2527.44 -> anterior circulation lesions.

2531.6 -> CTA was obtained.

2533.84 -> Basilar artery is severely stenosed at the origin.

2536.8 -> And both terminal vertebral arteries are

2538.96 -> practically non-existent.

2540.96 -> Here is the sagittal view.

2543.28 -> Incidentally, symptomatic large vessel

2545.04 -> atherosclerotic disease

2546.24 -> like in this patient

2547.52 -> is a common cause of posterior circulation infarction.

2553.12 -> So, our patient has terrible-looking vessels.

2555.52 -> NIH Stroke Scale score is 2.

2557.92 -> Is this a stable or unstable situation?

2561.36 -> Do we expect her to deteriorate despite medical therapy?

2566.56 -> Well,

2567.6 -> what about her CT perfusion?

2570.16 -> Her CT perfusion showed significant areas of Tmax over 6 seconds.

2574.96 -> And as expected,

2575.92 -> she deteriorated within 48 hours developing

2578.8 -> lethargy and bilateral leg and left arm weakness.

2581.84 -> And at that point,

2582.72 -> she underwent emergent mechanical thrombectomy and angioplasty.

2586.56 -> The vessels were so tortuous that a stent could not be placed.

2590.88 -> You may have noticed that there no areas of ischemic core

2593.6 -> (CBF less than 30).

2596.48 -> Don’t rely on the automatic CBF calculation in posterior circulation patients

2600.64 -> because it performs poorly.

2602.48 -> CBF less than 30 percent threshold was really validated for anterior circulation strokes.

2608.96 -> Post-intervention,

2609.84 -> there was still residual stenosis in the basilar,

2612.24 -> but at least the V4 segments of the vertebral arteries were open.

2619.76 -> Be wary of applying CT perfusion to posterior circulation strokes.

2624.08 -> Why?

2625.04 -> Well, mainly for 3 reasons.

2627.2 -> First,

2627.92 -> these patients were excluded from DAWN and DEFUSE 3 trials,

2630.72 -> and perfusion criteria for intervention in the posterior circulation

2633.68 -> are not well-defined.

2635.68 -> In fact,

2636.4 -> there is paucity of high-quality evidence

2638.24 -> that intervention is even effective.

2640.48 -> It’s a rare disease,

2641.76 -> often with poor outcomes without intervention,

2644.32 -> so experts are generally in favor of intervening in symptomatic patients

2648.64 -> rather than enrolling them into trials.

2652.4 -> Second,

2653.36 -> specificity suffers because of those pesky skull base artifacts.

2657.68 -> Here is what I mean.

2658.72 -> CTP erroneously diagnosed perfusion abnormalities in the posterior circulation

2663.28 -> in this patient

2664.64 -> with a right basal ganglia stroke.

2669.36 -> And the third reason is a technical one.

2672.08 -> Not every CT scanner can do whole brain (or volume) perfusion.

2678.56 -> Some only acquire two axial slabs that correspond to anterior circulation territories

2683.2 -> -- a protocol that is designed to fulfill the DAWN and DEFUSE criteria.

2687.04 -> You may have to specifically request the CT tech

2689.52 -> to lower the slabs

2691.12 -> if you want to image the posterior circulation in

2693.36 -> that type of scanner.

2699.04 -> Posterior circulation stroke is one example where the clinical picture may be confusing.

2703.84 -> Stroke mimics,

2704.96 -> like encephalopathy, seizures, hypoglycemia,

2708.16 -> are another challenge.

2709.84 -> Can we use CT perfusion to assess for stroke mimics?

2713.68 -> Maybe.

2715.76 -> Here is our last example.

2717.92 -> 63-year-old man

2719.28 -> with multiple vascular risk factors

2721.36 -> and history of remote renal transplant,

2723.36 -> immunosuppression

2724.56 -> and intracranial nocardia abscess, which was drained in 2020

2728.48 -> through a right craniotomy.

2730.64 -> He suddenly developed acute left hemiplegia, right gaze deviation, and left hemianopia.

2735.2 -> NIH Stroke Scale score is 13.

2738 -> Sounds like a right MCA syndrome, doesn’t it?

2741.44 -> Symptom onset was witnessed by his nurse.

2744.32 -> No problem.

2745.04 -> Let’s call a stroke code!

2747.2 -> Non-contrast head CT was normal.

2751.92 -> On the CTA, right middle cerebral and right cervical carotid artery are...

2756.56 -> open?!

2758.4 -> What!?

2761.04 -> Let’s run CT perfusion?

2765.28 -> No core.

2766.64 -> No penumbra.

2768.08 -> No mismatch.

2769.76 -> CT perfusion does not seem to be of much help here.

2773.28 -> Let’s look at the source images just in case.

2777.48 -> [Gasp]

2778.48 -> Do you see it?

2780.48 -> Tmax is symmetric.

2782 -> MTT is symmetric.

2783.84 -> But cerebral blood volume and cerebral blood flow is increased in the right hemisphere.

2789.36 -> Hyperemia can sometimes be seen after reperfusion,

2792.48 -> but this patient actually had a seizure.

2795.28 -> Based on the CTP findings,

2796.72 -> intravenous thrombolysis was withheld

2798.8 -> even though the patient technically qualified by clinical criteria.

2802.56 -> Video EEG showed multiple seizure episodes

2805.28 -> over the next 24 hours,

2806.96 -> and there was no ischemia on follow-up MRI.

2813.12 -> By now it must seem like CT perfusion is a superhero.

2816.56 -> There is nothing this modality can’t do.

2818.8 -> Can we just use it like the cardiologists use troponin?

2821.84 -> To rule out a stroke in a patient with acute neurological deficits?

2826.48 -> Not quite.

2828.88 -> This patient

2829.44 -> presented with acute right sensorimotor deficits.

2832.72 -> Stare at these image maps for a second.

2834.88 -> Do you see any abnormalities?

2836.64 -> I am even giving you this patient’s head CT as a clue.

2845.36 -> This patient has a left thalamocapsular lacunar infarction.

2850.56 -> Unfortunately, CT perfusion is terrible

2853.36 -> at identifying subcortical lacunar infarctions.

2856.08 -> I would argue that even the head CT picked up the stroke.

2859.36 -> Again,

2859.68 -> this is a stark reminder

2860.96 -> that even if CTP is completely normal,

2864.16 -> a patient may still have a stroke.

2866.96 -> Always examine CTP in the context of clinical history and other imaging.

2874.32 -> We are at the end of our tour through CT perfusion.

2877.76 -> If you had the patience to sit through the entire talk,

2880.24 -> it's time for a well-deserved summary.

2883.76 -> If a patient with acute stroke symptoms presents within 6 hours from last known well,

2888.32 -> CT perfusion is not necessary for intervention.

2891.68 -> But, it may be considered

2893.04 -> if time of onset is suspect,

2895.12 -> or there is a concern for medium vessel occlusion,

2897.6 -> or stroke mimic.

2900 -> In a patient presents beyond 6 hours from last known well,

2903.6 -> advanced imaging like CT perfusion is required

2906.56 -> to identify appropriate patients for intervention.

2910.08 -> Next,

2910.96 -> quickly assess the quality of study.

2913.12 -> I am talking about verifying appropriate vessel selection,

2916.32 -> appropriate contrast opacification,

2918.56 -> and minimal movement.

2921.84 -> If the study is inadequate,

2924.8 -> consider performing an MRI.

2929.2 -> If the study is adequate,

2930.96 -> then look for delay in flow on Tmax maps.

2935.92 -> If Tmax is normal,

2937.36 -> large or medium vessel occlusion is unlikely.

2940.24 -> Get a more detailed history and examination,

2942.48 -> consider stroke mimics,

2943.92 -> and take a look at CBF/CBV source maps.

2947.92 -> Again, the gold standard imaging here is MRI.

2954 -> Over 6-second delay

2955.36 -> is used as a surrogate for penumbra.

2957.52 -> An area of over 6 second delay is concerning,

2960.88 -> and we should review the CTA

2962.4 -> for the presence of a large vessel occlusion.

2964.88 -> If the intracranial vessel occlusion is absent,

2967.68 -> we need to explain the delay in flow

2969.2 -> by evaluating the cervical carotid.

2971.84 -> Note that we are talking about delays in the anterior circulation.

2975.6 -> If cervical carotid occlusion is present,

2978.8 -> consider that penumbra is being overestimated

2981.28 -> in a patient with chronic carotid occlusion.

2984 -> But before you call off the intervention,

2986.48 -> you need to correlate with clinical examination,

2988.96 -> non-contrast head CT

2990.64 -> and review the hypoperfusion ratio.

2993.2 -> Hypoperfusion ratio

2994.48 -> is the ratio between Tmax over 6 seconds

2997.28 -> and Tmax over 10 seconds thresholds.

2999.84 -> It may provide a clue

3000.96 -> that substantial amount of tissue

3003.04 -> is at an immediate risk of ischemia

3005.6 -> (like in this example).

3009.84 -> If cervical carotid occlusion is absent,

3012.88 -> consider artifactual result

3014.64 -> especially if the areas of flow delay are near the skull base

3017.76 -> or orbits.

3020.64 -> If proximal intracranial vessel occlusion is present,

3024.96 -> then assess ischemic core

3026.72 -> (area of cerebral blood flow less than 30%).

3029.52 -> Be very careful here

3031.2 -> since core heavily influences candidacy for thrombectomy.

3035.52 -> We start by comparing the CBF map to the non-contrast head CT.

3039.68 -> To be able to trust CT perfusion’s core estimation,

3042.48 -> obvious hypodensity on the head CT must be absent.

3045.68 -> Otherwise, you may be dealing with a rare situation

3048 -> where leptomeningeal collaterals were recruited late,

3050.88 -> cerebral blood flow had improved,

3052.72 -> but brain tissue already infarcted.

3055.28 -> So, the core may be underestimated.

3059.6 -> Ischemic core is over 70cc's

3061.76 -> generally increases the risk of reperfusion hemorrhage

3064.72 -> and makes intervention unfavorable.

3067.28 -> But before we finalize that conclusion,

3069.6 -> we need to double check that non-contrast head CT.

3074.08 -> If ASPECTS is low

3075.6 -> and a significant hypodensity is present,

3078.24 -> both CTP and non-contrast head CT agree that a large stroke has already occurred.

3083.52 -> Intervention should be avoided.

3086.16 -> If ASPECTS is high

3088.24 -> (there is no significant hypodensity),

3090.4 -> you may be dealing with a rare situation

3092.32 -> where core is overestimated

3094.4 -> in a patient who presents very early from stroke onset

3096.88 -> – perhaps even within the golden hour.

3100.08 -> Consider using the more stringent

3102 -> CBF less than 20% threshold

3104.32 -> to better estimate core

3105.92 -> in this situation.

3109.84 -> If the ischemic core is at most 70 cc’s,

3114.08 -> assess for penumbra/core mismatch.

3116.72 -> If a mismatch is present,

3118.32 -> the patient may be a candidate for reperfusion therapy.

3122.88 -> If the mismatch is at least 1.2 times

3125.44 -> (or 10 cc’s),

3126.8 -> TPA may be considered in a patient with a wake-up stroke

3129.76 -> within 9 hours from the midpoint of sleep

3132.4 -> (EXTEND trial criteria).

3134.64 -> If the mismatch is at least 1.8 times

3137.28 -> (or 15 cc’s),

3138.72 -> then mechanical thrombectomy may be considered

3141.04 -> per DEFUSE 3 criteria.

3146.4 -> And finally,

3147.52 -> if this process yields uncertain, confusing results,

3150.8 -> the ischemic core can always be evaluated more accurately

3154.64 -> with the gold standard diffusion weighted imaging MRI.

3167.12 -> And that’s all for today.

3168.64 -> As always,

3169.28 -> thank you for your kind attention.

3171.36 -> Until next time.

3173.36 -> Bye.

Source: https://www.youtube.com/watch?v=LQ80snVuELs